Advices on First Crossover Design (VituixCAD2)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
REW:
* No measurement sequence with automatic file naming: root filename + off-axis angle in degrees

"Export -> All measurements as text

Export the measured data for all loaded measurements as text files in a chosen directory. The file names use the measurement titles. The same settings dialog appears as for a single measurement, the settings are applied to all the files. Any note entered will appear in all the files, followed by any
notes from each individual measurement."


-while not automated:

Making Measurements

"Measurements are given a default name of the date and time at which they are made, a more appropriate name can be entered in the box at the top of the measurements panel.."
 
"Measurements are given a default name of the date and time at which they are made, a more appropriate name can be entered in the box at the top of the measurements panel.."

This is exactly the problem. It's a bit frustrating to modify names of for example 127 measurements which is possible and probable amount with horizontally asymmetrical 4-way. This becomes very frustrating if you need to repeat a single off-axis sequence due to noise or some bad mechanical/electrical/software setting in measurement setup.
Unnecessary manual work could lead to laziness and ignorance related to coverage of measurement data.

REW is one of the best choices for extremely poor people requiring freeware, and those who have endlessly time for non-productive tinkering. I suggest someone could inform author of REW that alternative naming system by root filename + running number with start, step and count (or end) is needed to get acceptance also for comprehensive speaker simulation. Pause timer and motorized turntable support would be useful too in addition to start measurement with space bar or some other single key. Until that REW stays as "not recommended software" at least in my messages and documentation.
 
Hi,


Unfortunately I have to say that I am now in possession of three vintage Scanspeak paperweights :(

A few month ago they all seem to work OK as I connected each to a channel of the amplifier, but today I was going to use the DATS V2 to measure them and they all confirm dead (10kohms) one of the three made a small click in the test but died also (this last one worked a few weeks ago)

I really don't understand how is that possible but I have to conclude that after being stored for years the dwindled and gave up in unison.


Even my multi-meter decided to give up today.

Post Mortem Pictures :


vjSLfzB.jpg


j5XZmvW.jpg
 
..alternative midrange:

8 ohm instead of 4 ohm at about 89 db, (which is a bit more eff. than the Scan Speak). Notably it has a much better *axial response higher in freq. - making the crossover to your planar much easier. It's excursion is nearly non-existent, but with a good high-pass filter this shouldn't be a problem, and it will lower non-linear distortion.

http://feleppa.com.au/pics/speaker_imgs/plots/FF85WK_FreqResp.png

http://feleppa.com.au/pics/speaker_imgs/plots/FF85WK_RawHarm.png


*also you can use the axial measurements on the plot above for basic yet accurate sim.s, perhaps asking Timothy for the files.. (..though note that condition for the spl is based on about a 1/3rd of a meter, so that if you going by 1 meter you'll need to correct.)
 
Last edited:
Do you get a lot of endorsement requests? :D

Not directly. Due to massive popularity of REW I have tried to give some technical support and instructions how to produce valid measurement data for crossover simulation. This has been so difficult and finally useless that I've given up. Very sorry about that.

Dual channel measurements with easily adjustable timing common for all measurements to export (as txt/frd), and easy and fast off-axis measurement sequence with automatic file naming by root filename and angle are very important features for crossover simulation with measurement data.

VituixCAD has separate tool for converting multiple impulse responses to frequency responses. Supported file formats are ARTA pir, CLIO mls*, wav and MLSSA ascii. Authors of ARTA and CLIO have published file format and supported my efforts to extend their products. REW mdat support is missing because I don't know the file format yet. REW has that part as integrated, but it's not so fast, easy and visual to use and remember as it should be imo.

I could add reading of mdat files (for time windowing, timing adjustment, FFT, smoothing and exporting to txt/frd) to VituixCAD if detailed documentation about file format is provided, and special database drivers are not needed. Version compatibility could cause some trouble I guess.
Off-axis measurement sequence can be implemented in measurement program only.

These features - especially off-axis sequence with suitable file naming has been okay in CLIO for "decades". Off-axis measurement and export sequences were added to ARTA few years ago after frustrating struggle with external tools such as AutoIT and ArtaRecorder (by me). Many thanks also to Dr. Weber who motivated Mateljan to do this.

I have some other wishes too, but those are not so mandatory in case mdat format is public. For example phase response should be visible while adjusting timing. Maybe excess group delay too i.e. single graph per tab is not the fasted and easiest UI for timing adjustments.
STI analysis is missing. That is more valuable for analyzing combination of speaker and room than simple clarity parameters calculated by decay.
Single channel gear such as USB mics and normalizing timing to IR peak are poison for crossover simulation with off-axis measurement data. Therefore public recommendation of Umik-1 without clear warning about bad timing issues does not look good. Acoustical reference exists, but it's not so nice and easy because mic position is not usually constant while measuring drivers of multi-way speaker.

I'm sure enough feature and change requests are recieved. So I'm not asking or wishing anything.
 
Last edited:
..alternative midrange:

8 ohm instead of 4 ohm at about 89 db, (which is a bit more eff. than the Scan Speak). Notably it has a much better *axial response higher in freq. - making the crossover to your planar much easier. It's excursion is nearly non-existent, but with a good high-pass filter this shouldn't be a problem, and it will lower non-linear distortion.

http://feleppa.com.au/pics/speaker_imgs/plots/FF85WK_FreqResp.png

http://feleppa.com.au/pics/speaker_imgs/plots/FF85WK_RawHarm.png


*also you can use the axial measurements on the plot above for basic yet accurate sim.s, perhaps asking Timothy for the files.. (..though note that condition for the spl is based on about a 1/3rd of a meter, so that if you going by 1 meter you'll need to correct.)


Hi ScottG,


Thank you for the suggestion of a replacement driver


I thinking the next course of action from my side after that misfortune and came with the following possibilities (I'll add yours also to the equation)


1- Repair the Scanspeak Drivers
I have contacted Scanspeak and they found me a repair place near me that I have contacted to evaluate the work and the cost of it.
basically it would cost close to $90 per driver (conversion from €) and tough I hate to discard the drivers I don't know if they are worth the cost and if the repair will respect the specifications.


2- I have looked over for mid replacement and came to the conclusion that the Seas Prestige MCA12RC - H1304 4.5" seemed to be good for a reasonable price (equivalent to the repair of a Scanspeak) I am particularly looking at the following Troels Gravesen baffle SEAS-3-Way-Classic


3- The suggested Fostex FF85WK
They are more difficult to find near me but they seem to be quiet less expensive than the Seas, I am confused with the specifications on Fostex website

The SPL parameter (dB not Sensitivity?) is 86.5dB/W(m) on the measurement from Timothy Feleppa it is a bit higher than 95dB/W(1/3m) most of the response and you say it is 89db.The Scanspeak where rated at 90db/2.85v/1m but in 4 ohms...I am lost.


Also the power rating 5w/15w (Rated/Max) is it not a disproportion with the other drivers ?


The off-axis response look quiet controlled but I have no experience to compare it with because most of Timothy measurement on mids seem's very good. Unfortunately I was not able to perform real life measurement with the Scanspeak, I only seen the simulated diffraction from VituixCAD with the front baffle of realistic dimensions from 3m or 10m, and Timothy do it with a large board of mdf from 30cm.


The only element I understand as a reference is the membrane size between the two drivers, Is it generally the greater the surface area the greater the dispersion?


Do you have the courage to keep enlightening me ? :D
 
Go by the graph in the data-sheet for Fostex for overall eff., and in your bandwidth it's mostly 89 db.

Subtract 3db for a 4 ohm driver (from its graph at 1 watt/meter) to get it's 8 ohm eff. for comparison.

The Fostex's power rating, like it's non-linear distortion, is almost meaningless in a design with a higher freq. high-pass filter (..and we are talking about something near 400 Hz with a fairly "aggressive" filter). It's also something that you should be able to model as far as power/compression (with that filter).

The Scan Speak driver isn't as linear as the Fostex near the intended crossover to your planar. I *think* it's dispersion isn't nearly as good either. The smaller the driver diameter the GREATER/wider the dispersion. Your Planar will almost be "omni" (..err "widest"?) on its low-end where crossing over to the midrange - you want that near "omni" dispersion for the top-end of your midrange where its low-pass will be. This means as you move side-to-side from the loudspeaker (horizontal axis) - you won't be getting suck-out (..a "dip" in freq. response).

As far as Troel's description of the Seas driver..

That's 1. less efficient, and 2. has more than double the mms. - of the Fostex.

In other words: the Fostex should be considerably better with his favorable discription of the Seas midrange. Note that the Fostex is also more similar to your planar mechanically (..more efficient and lower mms). I'm assuming you chose the Planar because of it's apparent "speed"/clarity; you do want similar character for your midrange, yes?

It's up to you, but in your price range and your prospective design - the Fostex is what I would choose any day over the Scan Speak or the Seas.


Also consider your mid-bass driver in relation to this. Lower mass, higher eff. is generally going to produce a better result, and make it easier as far as baffle step loss is considered. Worry less about fs for that selection, because higher eff. drivers are going to be higher in fs., but that's where lower freq. tuned ports (bass reflex) help-out. It's going to be a larger volume box though..
 
Last edited:
..so French or German suppliers?

French seems better for Fostex:

Haut-parleur large-bande Fostex FF85WK, 8 ohm, 83 x 83 mm

Fostex FF85WK - Mega Audio
https://www.megaaudio.de/Downloads/Preislisten/Fostex-Hi-Fi-VK-Preisliste.pdf


..and assuming you want to keep it to an 8 inch max diameter for the midbass (German supplier):

Same basic driver, different prices (..non-linear is a bit worse for odd-order harmonics with the less expensive driver, and 2nd is higher in the lower midrange - though you'd be hard-pressed to actually notice 1% 2nd order):

https://www.intertechnik.de/media/1381683_88463.pdf

https://www.parts-express.com/pedocs/specs/295-345-Dayton-Audio-PM220-8-Specifications.pdf

https://www.intertechnik.de/shop/la...axis/gradient-axis-axp-08_1768,de,6800,145293

https://www.intertechnik.de/shop/la...tbandlautsprecher/pm220-8_1768,de,7055,150445


The closest "analog" you'll find on Troel's site is the Supravox 215 GMF and the subjective description at the bottom. Truth be told though, he didn't use it correctly. Cabinet volume and vent freq. were to limited. He should have gone lower (vent freq.) with a larger volume enclosure (..though it does limit spl), plus the 215 GMF really isn't a driver designed for freq.s much past 500 Hz.

http://www.troelsgravesen.dk/Supravox215GMF.htm
 
Last edited:
Go by the graph in the data-sheet for Fostex for overall eff., and in your bandwidth it's mostly 89 db.

Subtract 3db for a 4 ohm driver (from its graph at 1 watt/meter) to get it's 8 ohm eff. for comparison.

The Fostex's power rating, like it's non-linear distortion, is almost meaningless in a design with a higher freq. high-pass filter (..and we are talking about something near 400 Hz with a fairly "aggressive" filter). It's also something that you should be able to model as far as power/compression (with that filter).

The Scan Speak driver isn't as linear as the Fostex near the intended crossover to your planar. I *think* it's dispersion isn't nearly as good either. The smaller the driver diameter the GREATER/wider the dispersion. Your Planar will almost be "omni" (..err "widest"?) on its low-end where crossing over to the midrange - you want that near "omni" dispersion for the top-end of your midrange where its low-pass will be. This means as you move side-to-side from the loudspeaker (horizontal axis) - you won't be getting suck-out (..a "dip" in freq. response).

As far as Troel's description of the Seas driver..

That's 1. less efficient, and 2. has more than double the mms. - of the Fostex.

In other words: the Fostex should be considerably better with his favorable discription of the Seas midrange. Note that the Fostex is also more similar to your planar mechanically (..more efficient and lower mms). I'm assuming you chose the Planar because of it's apparent "speed"/clarity; you do want similar character for your midrange, yes?

It's up to you, but in your price range and your prospective design - the Fostex is what I would choose any day over the Scan Speak or the Seas.


Also consider your mid-bass driver in relation to this. Lower mass, higher eff. is generally going to produce a better result, and make it easier as far as baffle step loss is considered. Worry less about fs for that selection, because higher eff. drivers are going to be higher in fs., but that's where lower freq. tuned ports (bass reflex) help-out. It's going to be a larger volume box though..


Thanks for elaborating on this, It didn't immediately make sense to me the link between the mid & tweet.



I still don't understand why smaller is less directional but that is to expected, i'll read about it when I have some time.


so less moving mass means better control over the excursion and then also a more precise and detailed sound I guess, less distortion.


Fostex it will be then
 
..so French or German suppliers?

French seems better for Fostex:

Haut-parleur large-bande Fostex FF85WK, 8 ohm, 83 x 83 mm

Fostex FF85WK - Mega Audio
https://www.megaaudio.de/Downloads/Preislisten/Fostex-Hi-Fi-VK-Preisliste.pdf


..and assuming you want to keep it to an 8 inch max diameter for the midbass (German supplier):

Same basic driver, different prices (..non-linear is a bit worse for odd-order harmonics with the less expensive driver, and 2nd is higher in the lower midrange - though you'd be hard-pressed to actually notice 1% 2nd order):

https://www.intertechnik.de/media/1381683_88463.pdf

https://www.parts-express.com/pedocs/specs/295-345-Dayton-Audio-PM220-8-Specifications.pdf

https://www.intertechnik.de/shop/la...axis/gradient-axis-axp-08_1768,de,6800,145293

https://www.intertechnik.de/shop/la...tbandlautsprecher/pm220-8_1768,de,7055,150445


The closest "analog" you'll find on Troel's site is the Supravox 215 GMF and the subjective description at the bottom. Truth be told though, he didn't use it correctly. Cabinet volume and vent freq. were to limited. He should have gone lower (vent freq.) with a larger volume enclosure (..though it does limit spl), plus the 215 GMF really isn't a driver designed for freq.s much past 500 Hz.

http://www.troelsgravesen.dk/Supravox215GMF.htm


Thanks for the Fostex Links, the french one I had found, I never ordered anything there, I'll check more thoroughly.


You keep surprising me, never heard of that Gradient Acoustics, are they Chinese made ? There is no support link or anything describing the company, have you any experience with the company and their drivers ?


Besides that the two suggested drivers are wide/full range, why not use a driver that is more oriented toward the lower frequencies, is it because crossover with the Fostex?


The AXP-08 does not provide excursion range or any power specs maybe I have missed it, the MMS is also very light it is mandatory to go to the higher frequencies I suppose, VAS is low so smaller box I guess, and it is very affordable.



The PM220-8 is more sensitive with a even lower mass it seem very similar in concept to the AXP-08


How would you describe the possible end result of using the LS Plan, and the suggested drivers in a three way baffle ? I have have some difficulties to imagine. What would they be best suited to listen to ?



This may not help you but I possess at the moment a Koda KD-261 or Onkyo A-7070 Intergrated Amplifiers with a pair of Cabass Jersey 220 4ohms



I was thinking of going toward an Hypex N-Core amplifier or an Emotiva, but the latest are quiet more expensive in my country than in the US, but I am not even sure I really need to change.


Thanks
 
Last edited:
You keep surprising me, never heard of that Gradient Acoustics, are they Chinese made ? There is no support link or anything describing the company, have you any experience with the company and their drivers ?


Besides that the two suggested drivers are wide/full range, why not use a driver that is more oriented toward the lower frequencies, is it because crossover with the Fostex?


The AXP-08 does not provide excursion range or any power specs maybe I have missed it, the MMS is also very light it is mandatory to go to the higher frequencies I suppose, VAS is low so smaller box I guess, and it is very affordable.



The PM220-8 is more sensitive with a even lower mass it seem very similar in concept to the AXP-08


How would you describe the possible end result..

Gradient is just an aggregate supplier like Jantzen and they have been around for a while.. and most drivers are sourced from China or Indonesia these days - it's not good or bad: it just "is". I'd almost guarantee that the drivers were produced in the same factory. The listed xmax is probably identical, though in point of fact the major limit isn't its force curve VC position, but rather the spider on these drivers (..compliance), which is tiny for a typical 10% distortion rating (..but so are a lot of drivers with far higher rated xmax). It's not something I'd be worried about UNLESS I wanted very loud continuous spl. ex. I would NOT use this driver as a single woofer driver in a venue-placed sound reinforcement loudspeaker (..though it would make a fine midrange there).

Note that Intertechnik has a design with it and a Seas coaxial, used in a similar fashion to what I'm suggesting.

Club23 | Speakers Intertechnik - Shop - Shop



Chances are the Gradient spec. isn't super accurate for any given driver (..because its spider-limitation is quite dependent on how much the driver is "broken-in" and sample-to-sample variation).. BUT using Gradient's/Interetechnik's T/S param.s in a typical box calculator.

Intertechnik's measurements (scroll down to 8" version):

Two new loudspeaker drivers - November 2013 - Loudspeakermagazine 2013 | Loudspeakerbuilding


-around 70 liters, tuning freq. of about 26 Hz. Probably a double rectangular Vent(s) of about 1.5 inches by 4 inches and about 17 inches long (each).

Assuming it's about 94 db (Anechoic) at 180 Hz, that net's you around 89 db for typical baffle-step loss (independent of any floor-gain). I'm figuring about 5db of baffle-step loss. At 100 Hz you'll be down more than 1db.. BUT you'll also have more floor-gain here. That still probably puts you a few db high from your target, but not a lot depending on the design, and a slight downward sloping "tilt" from 100 Hz - 20 kHz is often preferred (..particularly where the response is less directional or "wider" between 1-8 kHz).

This won't ever go super loud without distortion or power compression, but it will go loud enough for most domestic listening. Use any amplifier you want with it (..provided the amp can handle the resulting Impedance & Phase.) Any type of music: even compressed Hard Rock reasonably loud.



As far as value: it's a "no-brainer" IMO, I'd go for the Gradient vs. the Dayton in this instance. x3 price provides exceedingly little performance gain. :eek:
 
Last edited:
Yes you are right Chinese isn't bad per se it's the quality control of specifications that matter

I always linked in my mind that the excursion was the main factor for bass reproduction seeing all those sub-woofers equipped with giant butyl rubber surrounds, but this only matter for very low frequencies and SPL performances.

Tough I believe you I am still wary of paper or tissue surround due to the many crappy sounding drivers that where equipping "value" products years ago, or those foam surrounds that dry and breaks after some time.

I don't listen very loud, it's difficult to describe but since I bought that Beyer Dynamic DT990 headphones I feel like the sound out of my speakers aren't as satisfying, it's sound kind of confused, I'd like the sound to be detailed (Airy?) and dynamic (percussive?) can go low with big drums, don't shriek with flutes, and clear voices. (only that;))

I will do a bit of simulation and tinkering based on your recommendations LSplan+Fostex+Gradient, do you think there is any advantage to go to the AXP-10 if I can fit them ?


BTW The AXP speakers surround remind me a bit of this other dead speaker I still did not get rid off I don't know why (hoarding disorder?):


WiXRh4U.jpg
tDr4adF.jpg




Thanks
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.